Bank Bottles

4000 PSI was the minimum bank pressure I had targeted for the air banks of this fill station.  This was due not for bank-capacity reasons, but that my PST tanks are all high pressure (3445 PSI) versions.  4,000 PSI banks would allow for several fills  (depending on the banks' capacity) before having to run a compressor to top the banks off.  Smaller banks deplete more often during larger fill jobs and become inconvenient when time is pressing.

Bank bottles are not commonly made for just 4,000 PSI: the general choices in the neighborhood are 3000, 3600, and 4500 PSI. Prices rise dramatically for the 4500 PSI bottles over the others by  almost $150-$400 each for new bottles.  I refused to invest that much for a set of new bank bottles, so I began looking for used and/or refurbished bottles.

Royal Pain

The first reasonable company I contacted (on a tip from Ted Green) that sold used cylinders was Queen Cylinder of Ohio.  Queen was selling refurbished 4500 PSI bottles with the buyer's choice of a new CGA valve, choice of new paint, and fresh hydro for only $225.  Freight for ten bottles to eastern VA was $270.70.  The price of their new 4500 PSI bottles at the time was $350.  Ten refurbished bottles was a bit more than I wanted to invest in at the time.  As it turned out, the banks for this fill station ultimately ended up being the largest investment in the entire project.  I located some other locals who wanted to get in on the purchase & shipping deal for a few bottles of their own, so this spread the costs around for everyone that was interested.  Around mid-year 2005, I was finally prepared to pull the trigger on a purchase and contacted Queen again.  Unfortunately, they had raised their prices by 50% in the previous few months since I had last contacted them. Their excuse for the sudden increase was that the bottles had suddenly become very popular with the paint-ball crowds.  Refusing to pay the extra premium, I went looking elsewhere for better pricing.

The Lucky Break
A fellow diver offered another tip that led me to Air Power International, located next to Philadelphia:   They had a good number of "refurbished" 3600 PSI bottles being sold at the time for $100 each in quantities of 10 or more, or $125 individually.  Many of these bottles were manufactured in the 70's & 80's and were close to (or just beyond) their 2'nd or 3'rd ten-year hydro test-date.  These bottles looked at least as robust as the ones that you rent from your local gas suppliers: ones that have bounced around on the local trucks for a 30-60 years at a time, subjected to rapid fill-speeds at the industrial gas plants.  The valves on these were all CGA-347 (3000 -5500 PSI service) and had burst disks rated for 6200 PSI.  My concern over the bank bottles being out of hydro was not as much a factor as long as the burst disk did its job and failed before the bottle did.  But for peace of mind, I checked with local hydro-testing companies and hydros for these generally cost $25-$35 per bottle.  In conversation with Steve Furst of Air Power International, he confirmed that these older bank bottles (like most) were made for a very long and abusive service life and consequently over-made as safety factor against that.

3600 PSI is only 155 PSI more than my HP steel SCUBA tanks required for a full fill at 3445 PSI.  A large number of bank bottles would be needed in order to be able to fill just one set before having to run the compressor.  If you are comfortable with the controversial idea of a 10% overfill, this turns a 3600 PSI bottle (3600 + 10%) into a  3960 PSI bottle, which is very close to the 4,000 PSI capacity sought in my original plans.  Overfilling tactic triples the amount of air a single bank bottle could provide to top off HP steel tanks.  10% over-fills are very likely still within reasonable safety margins.  This decision led to purchasing  some of these 3600 bank bottles with the intent to keep them pumped up to the 4000 PSI originally planned for. 

Most of these bottles came from fire stations that are upgrading their older SCBA systems (3000-3500 PSI) to higher-pressure (4000-5500) SCBA systems. In the wake of 9/11, officials in New York City and other municipalities suddenly experienced the problem of having to return to home stations to recharge their SCBA bottles, instead of right on the emergency vehicles on-site.  API has been designing and installing special Mako SCBA compressors that mount on emergency vehicles and are driven off hydraulics or generators.  After these upgrades, most stations have little use for the lower-pressure SCBA systems in house or bank bottles.  API has been buying these older systems as trade-in value.  Steve Furst said API maintained and regularly serviced (including all filters) the systems from which he obtained these bottles, so he knew they were not contaminated or abused by the compressors that filled them.  In the few bottles that I later opened up, bare metal was still showing in most places, with a few streaks of brown on some, probably from water left over from their last hydros.

Morphing Designs

10 bank bottles was more than the 6 bottles than my original plan required.  But with reasonably-priced bottles available, further possibilities open up.  Having a limited set of doubles and deco stages at the time, I would often have to dump some of the mixes in order to remix them into a new blend.  With a helium/oxygen analyzer and a Haskel, there was just no compelling reason to dump perfectly good mix vice recycling into later mixes.  So the two extra bottles were dedicated to hold leftover nitrox and trimix.  The bank layout then became  2 air banks of 3 bottles each, and one air bank of 2 bottles.

My oxygen and helium supply bottles from  a local gas supplier cost me $65 a year to lease.  This supplier confided to my buddy, who also leases bottles there,that they make a lot of their profit from bottle leases.  This supplier won't allow or condone customer-owned bottles at all, as you can guess.  The cost for a 282cf bottle of oxygen was $13, so it was worthwhile to continue dealing with them, factoring the bottle lease into the yearly consumption rates.  However, these costs caused me to consider alternatives.

Air Power International was selling used 2400 PSI bank bottles for $75 each at the time, which led to the idea of adding an oxygen cascade to the project.  Employing a Haskel, I could make one trip to pick up 3-6 oxygen bottles from the gas supplier, pump my oxygen cascade full, and take them all back after a day or two.  Outside of an extra trip, doing this traded a yearly lease for just a couple bottles into a daily lease for a lot of bottles. Daily-lease costs are <$0.20 per bottle, or somewhere close to dividing the yearly lease rate by 365 days.  Elimination of the $65 annual-lease cost per bottle would actually pay for privately-owned cascade bottles after their first year of ownership. 

I always had to run the Haskel in order to get service-pressure (3,000 PSI) fills  into my 100% deco bottles.  So after further thought, the 2400-PSI bottles seemed like a waste when 3600 PSI bottles were available for 30% more in cost..  They would allow banking oxygen up to 3600 PSI, instead of the 2200-2400 PSI that normally came in a supply bottles.   This would allow for a true, cascaded, O2-on-tap design and reduces having to watch over a Haskel every time O2 top-offs are needed.  Of course, 3600 PSI was well above the pressures I intended to boost oxygen to.

The desire for an additional 4 bottles for the oxygen cascade also led to consideration for a helium cascade for the same reasons.  This culminated in planning for an additional pair of bank bottles for a helium cascade, which could be filled close to 4000 PSI .  I made a trip up to API and brought back ten more bottles (exercising the 10-bottle bulk discount), which was all that API had available at the time.  While picking the bottles up, I was offered four HC-4500's they had on hand for $150 each.  The price was very attractive.  Thinking quick, these bottles would actually serve well as a fourth, highest-pressure air bank, fully rated for the maximum pressure planned.  These extra bottles suddenly became part of the project and bought them on the spot.  Steve Furst pointed out that these HC-4500's were not designed as robust as the other bottles.  I just lifting them up by hand, I could tell that they were much lighter than the other 3600-PSI bottles I had bought.  HC-4500's are not recommended for pressures higher than their rating.  But this was OK since I did not feel like pushing my RIX compressor that high already expecting increased ring-wear pumping up to 4000 PSI alone.

CGA Valve Care
A few of the Sherwood CGA 347 valves on the bottles I bought were leaking a little and difficult to tighten down the knob enough in order to stop the leakage.  I replaced the soft seats and some stem packings on these.  The Lower Plug & Seat Assembly only cost: $2.16 each and the Teflon packing cost $0.71 each



CGA valves come apart very similar to a SCUBA tank valve.  The valves on the 4 designated oxygen-cascade bottles were completely disassembled and oxygen-cleaned.  See the section further below on oxygen-cleaning for the cleaning method I used.

I also oxygen-cleaned the four bottles intended for the oxygen-cascade.  Removing the valves is not as difficult as some have expressed.  I could have wrenched the valves off  with a 15" pipe-wrench if wanted to.  But instead of marking up the sides of the soft brass CGA valves with a pipe wrench, I happened to have a very-large adjustable wrench that suited better.  Removing the valves did not require a "cheater bar" at all.

To get the CGA valves off, I made up a jig to hold the tank from spinning.   I  fastened one end of some 1/2" all-thread rod to a 2x4 board.  To get the bend close to perfect, I put a bottle next to the all-thread and bent the all-thread around the bottle until the free end touched the 2x4.  This initial bend sets the general shape of the bottle into the all-thread and also leaves the far side as straight as possible for where it has to fit through its own hole in board. For a 9.5" diameter bottle, the holes were 9" inside-edge to inside-edge. The rubber strip in the picture was just something I had laying around that worked well enough to protect the finish of the bottle from the threads on the rod. Bicycle tire would have suited just as well.

Next, I unfastened the all-thread off the board and used the bottle to wrap the all-thread around into a complete U.  Next, I inserted both ends  of the all-thread into the board and put on the bolts.  I recommend large metal washers to prevent the bolts getting sucked into the wood when tightening them down.  I cut off the excess length from the all thread but making sure it left a couple inches of clearance when sliding a bottle into the loop.  I dressed up rough edges of the cut with a metal file.

Also pictured is the adjustable-wrench I used to unscrew and tighten the CGA valves.



This is the cylinder secured in the jig.  The all-thread nuts shouldn't have to be tightened down too far to provide enough grip on the tank.  Note the rubber band placed between the all-thread and tank-surface.


I used the following method to wash out the bottles. It was taught to me at a local dive shop for oxygen-cleaning regular SCUBA cylinders.  Please do not accept this method as the only or best way to do it.  Other methods may be more easier and/or more effective.

I filled the bottles half-full of very-hot water.  This amount allows for the water to move with some speed when sloshed from end to end, and also expose the sides to a good amount of water when it rushes by.  For detergent, I add a fluid ounce of Simple Green concentrate.  To agitate the water inside, I set the tank on a 4x4 block and rock it back and forth like a see-saw.  I waited to hear/feel all the water slosh against one end before tipping the tank to the other end.  A wet rag on top of the 4x4 helped prevent the tank from sliding on the wood as it rocked from end to end.  After 4-5 sloshing cycles in one position, I would rotate the tank about 30 degrees and do another 4-5 sloshing cycles.  In this manner, I averaged about 1/4 turn every minute or so.  I spent about 4 minutes to make a complete rotation of the tank, which exposed all the inside surfaces to an equal amount of washing action. Rocking the tank for a while may cause your wrists to ache, especially  after doing several bottles!  I would make a second or third complete washing-rotation before I felt ready to drain the tank. In all, this process took about 8 -12 minutes.

Next, I would fill the tank half full with clean (hot preferred) water for the rinsing cycle, which was basically a repeat of the method already mentioned above.  After finishing the initial rinse, it was not uncommon to see a few soap suds float out with the last drops of rinse water.  I would then follow up with a second rinsing cycle.

An easier (and more water-conservative) way to rinse any tank is to make a fitting on the end of a solid pipe that could fit inside the tank opening and reach to the furthest end.  A spray-nozzle of some sort on the end up the pipe would make an umbrella-shaped spray inside an inverted tank to effectively rinse all the walls at once as the sprayer is withdrawn from the tank.. Regular garden-hose fittings are too wide to fit through an NGT tank-opening, or they might make a good nozzle


This drying stand (has a large hole in the cross-piece to secure the tank's neck) and pipe was used to blast-dry the bottles after they were washed.  I set the tank upside down on the stand and insert the copper tube all the way into the tank.  The pipe was used to deliver air to the top of the tank, and force the water droplets inside down and out the neck.  I have a shorter, smaller version that I use for drying SCUBA tanks.


The end on the drying tube pictured above is just a piece of half-inch copper tubing with a flare fitting at the bottom.  A brass elbow, and one of my favorite Parker quick-disconnects rounds out the materials for this.  I use the Parker QD's on all my whip hoses and other whip parts to allow me to connect just about anything to anything I need.   See the QD section below for more information on these handy parts.  Of course, everything shown here is oxygen-cleaned and stored in plastic bags when not in use to keep ambient contamination away as much as possible.


This crude stand held the bottles upright while blasting them dry.  The simple tank washing/drying stand pictured in the Oxyhacker's Companion is really the way I would recommend inverting the bottles with.  Heavier boards would be needed for the heavier bank bottles though.

I place some material underneath the tank's opening to prevent dirt and dust from flying everywhere when the SCUBA air supply valve is opened all the way.  For reference, this force and volume of the escaping air will actually lift aluminum tanks right off a stand..

An AL-80's worth of air should be more than sufficient to dry any tank.  To start, I open the SCUBA valve just enough to release a full blast of air for about 2-4 seconds.  This initial blast forces out the majority of water droplets clinging to the surface inside.  Next, I close the valve down quickly until it is just cracked enough to take about 4 minutes to empty the AL-80.  This seems to give any remaining moisture in the cylinder time to evaporate in the super-dry air flowing past it.  A tank still warm from a hot-water washing helps speed the drying up.  I didn't notice any flash-rusting on most of the bottles after following this method, but it is a possibility if one waits too long to dry the tank after rinsing it out.



To screw the CGA valves back into the bottles, there is guidance in the Technical Specifications PDF on Sherwood's site that describes the special tightening technique used for the National Gas Taper (NGT)-threaded CGA valves.  If you're not familiar with the NGT tightening-method, you should read this before attempting to install the valve. You probably WILL need a "cheater bar" to get the valves back in tight enough.

As shown here, it can take quite a bit of leverage to get the valves tightened back to NGT specifications.  A piece of pipe is being used as a cheater bar here.


Once the cheater bar is nearly laying on the ground, I picked up the jig and continued to torque the valve as shown here.  This method allows for the awkward angles that would otherwise be encountered by just moving the wrench a half-turn on the valve



1 - Introduction     3 - Fittings    4 - Tubing     5 - Valves     6 - Manifolds, whips, gauges,O2-cleaning     7 - Results & Pictures